Joty—The Order of Formation of Silicates in Igneous Rocks. 301 
Calcium Oxide.—This can be crystallized by sublimation in the 
oxy-hydrogen flame, covering the lime cylinders with a thin trans- 
parent film and brilliant crystals (Proc. Royal Dublin Society, 
vol. vi., p. 225). M. Moissan reports that it cools from liquidity 
to a crystallized solid. It possesses a very high melting point, 
which is lowered by addition of Al,Os. 
Magnesium Oxide.—Much like calcium oxide, but with a higher 
melting point: gives in the electric furnace large, transparent 
erystals out of the liquid, often several millimetres in length. 
Sesqui-oxide of Iron.—F uses rapidly, losing some oxygen, and 
pasing into the ferroso-ferric oxide, Fe;O,, which remains partly 
erystallized. It combines vigorously with the material of the 
furnace. 
Potassium Oxide and Sodium Oxide.—These are known to become 
liquids at a full-red heat. Further information I have not obtained. 
Now, if we assume that the properties of the first three of 
these bodies are in any marked degree additive, there is evidently 
a full explanation of the apparent abnormality in the order of 
erystallization of many silicates. ‘The silica enters as an influence 
retarding crystallization and prolonging the viscous properties 
downwards in the scale of temperature. CaO, MgO, A1,O;, on 
the other hand, are ecrystallizers at high temperature, and influence 
the molecule accordingly. 
That some such additive result is to be expected follows, 
perhaps, from the nature of the silicates, which, as Prof. Mendelieff 
observes, partake greatly of the nature of alloys.’ He adduces the 
fact in support of this, that their specific volumes are additive. 
Thus, the volume of the silicate orthoclase can be closely calculated 
from the volumes of the constituents—A1,O;, K.O, 6SiO,; and 
also the fact that the chemical union of these oxides is feeble, 
water, H,CO;, and such weak agencies, being able to gradually 
break them up.’ ‘The properties of glasses, too, are in many 
1 It is noteworthy, in this connexion, that the solidifying points of alloys are only 
exceptionally below or above those of the constituents. These exceptions are well- 
known. Such curves as Gautier’s, while showing that rarely is the solidifying point 
influenced directly proportionately to the percentages of ingredients, reveal that the 
introduction and increase of a constituent of high melting point involves, in general, 
a rise in solidifying point. 
2 « Principles of Chemistry.’’ London § New York, vol. u., p. 118. 
